Battery replacements for Valve Radios

Hi Ivan,

Think this my answer your questions, and more!

Keith

Click to see full size image

Click to see full size image

Click to see full size image

Perfect. Thanks Keith!

Just what I'm looking for.

I calculate that a pair of Alkaline D cells is about x4 the capacity of the slightly larger Zinc Carbon that were in a A35, if they were longer than D cells, i.e. about 75mm long

30 x CR3032 is now cheaper and 20% life and 1/3rd size of PP3 Alkaline. Two stacks with 1N4148 splitter diodes is  viable to give 480mAH to 560mAH depending on quality with about 1/2 or less voltage droop with age.

Alkaline, Zinc Carbon (original) and 1x CR2032
So multiply 1st 2 graphs by 10 and last graph x 30  

There are numerous issues with 10 x PP3 NiMH in terms of over discharge of one of the 60 cells in the 10 packs/Batteries and how to charge.

I'm going to do a 90V pack with LT and HT sockets and 6 x C size NiMH and Boost inverter. A flying lead will connect to a A35 pack with 2 x D cells. Current flow on LT is detected to power up the 90V boost inverter which must work from 5.4V to  9.5V. Recharge from 12V (built in smart charger so any source 11V to 16V to charge or run).

Sounds good.

I would humbly suggest that Keith's data should be PDFed and put in The Library.

- Joe

Tis Done, http://vintagetvandradio.myfreeforum.org/about263.html

Chris

Thank you.

Grid Bias batteries are something I'm yet to encounter.

- Joe

Grid bias battery feed from 3 x CR2032 should last YEARS. But designers quickly learnt tricks.

Like series filaments on direct valves.

My MkI prototype case made of paper from laser print pritt-sticked to card.

Click to see full size image

Cut piece of plain PCB size of box and poke 2 holes. Enlarge with scissor till plug just fits . Cut isolation with Stanley knife by two score and peel out strip.

Make two connectors out of coffee tin.

Click to see full size image

Click to see full size image

Click to see full size image

Velcro to close lid

Click to see full size image

battery holders made from 2 x pieces of coffee tin and soldered to PCB.

Piece of plastic over PCB to insulate from battery cases

Centre connector has plastic barrier between pins and glue gunned and then taped.

First attempt at an A35 box that uses replaceable Alkaline D cells. Should be about 16,000mAH with good Alkalines.

Use from same packet and check they match before fitting. Don't use Rechargeable cells in this holder.

If DL96 is 50mA and the others are all 25mA = 125mA for standard Dx96 line up.  About 130+ hours with decent quality Alkaline vs about 30 to 50hrs for original?

my son said:

"So you had an old Battery box?"

"No, it would be blue if there was any inkjet ink"

He then thought...
"everything wasn't B&W back then "  (Allusion to Calvin & Hobbs)

You're certainly a fast worker.  It looks good too.

- Joe

I'm off to buy a load of CR2032 now  for the HT pack

http://shop.ebay.ie/i.html?_nkw=CR2032& ... 01&_ftrv=1
or
http://shop.ebay.co.uk/i.html?_nkw=CR20 ... 01&_ftrv=1

20mm diameter, 3.2mm thick Nominal 3V, 220mAH to 280mAH depending on quality / make.

Easily stacked in a tube.

How bad could these be?
http://www.ebay.co.uk/itm/100-Lithium-3 ... 0354872624

90 as 3 x stacks of 30 + 3 x 1N4148?

Under £15 for more than 600mAH ( more 3x capacity of Alkaline or NiMH 10 x PP3)

Well the replaceable 2x D cell "A35"  works on the Invicta 29

Much less interference on Battery.

I used croc clips onto a stack of 12 x CR2032 and  4 x PP3 in series. HT current was 5mA and gradually rose to 6.5mA, then dropped to 6.3mA

So 660mAH (90 x CR2032 in three "stacks") would be over 100 Hrs. A set of ten Alkaline PP3 would be about 32 hrs HT.

Anyone ever fitted 10 rows hexagonal close packed of 6  x AAA cells in a B126 box?

I think it fits, but may not leave space for socket!

Also figuring how to make them replaceable is another story. I've lots of empty Lidl coffee tins.

The B126 box will take 5 x sets of 30 x CR2023. Four columns of 30 and 2 of 15 leaving plenty space for connector. That's 1100mAH (175 Hrs)  vs  200mA  (32 hrs) for PP3s

The AAA Alkaline cells are about 800mAH ( 125 hrs approx)

well here is i pic of an original

and here is my ht supply

and here fitted into old battery boxes

and finaly the circuit of the ht modual

rob t

Very interesting, Rob, especially the shot of the original battery guts.

- Joe

Nice pics.

The following all approximate of top of my head

3 x Alkaline AA = nominal 5V approx  @ 2500mAH

Say O/P = 90V at 6.0mA

so 5V @ 100% efficiency = 108mA

Let's say 90% efficiency (generous), so 120mA

2500/ 120 = 21 hours approx

If you use 6 x C cells = 10V nominal and 4000mAH

so 60mA (as twice the volts)
and 4000/60 = 67 hours approx

I'll build the LT1073 based switch. But I want to detect LT current to enable/disable

I've been looking at LT1083 of and on...

These have a load simulating typical 4 valve Dx96 type radio:

Click to see full size image

This is based on application note. It's very close to the 50V limit of the LT1083 internal transistor switch and needs 100V 1uF Ceramics.

Also THREE posh high speed Diodes. 1N400x family won't work
Seems a nice idea but not good above about 0.5mA

Click to see full size image

So what if we use an external switch? We need at least 6 x NiMH to have enough volts at end point (0.9V) to operate a FET. A bipolar transistor isn't great for this. The low switching frequency mitigates high gate capacitance

Click to see full size image

Normally there would be a Schottky diode and 2N3906 at the gate of FET for faster turn off. But the IRF510 is lowish capacitance and the extra diode would lose 0.5V of input switching level.

Click to see full size image

Much better performance. Probably be better to have a slightly higher voltage rated FET. I'll likely put 1K in series of the 2.7K on Voltage feedback to reduce output a little.

The low value of capacitor on diode reduces peak of FET drain voltage.
Improved O/P filter. The 100uH has only 6.5mA so a "resistor" type one will do. The Switching inductor needs to be much lower resistance as there is peak current of over 1A.

Next is to only enable the HT supply if there is LT current. A fly lead to the LT supply and socket for LT beside the HT socket on the "fake" B126.

6 x C size NiMH and the electronics do I think fit.

sounds about right useing 6 2500 mah rechargables gives about 40 hours use on the ones i do now i remove the 5 volt regulator because it serves no purpose
would it be posible to make the circuit on vero board? as the max chips are cheap and the moduals now cost £10.75
rob t

Farnells do the MAX1771 ICs at over a fiver plus VAT.

The data is here:-

http://www.farnell.com/datasheets/74526.pdf

- Joe

The LT1073 is about 40p to £1. There are Maxim and Motorola equivalents.

A suitable MOSFET and high speed / ultrafast diode are cheap too. I don't do "dear".

I'll be building it on veroboard and cased for screening in "coffee tin" (light weight scissors cut it and it solders with 25W iron easily).

There will be option for 12V jack and charging circuit for NiMH (which are similar capacity at LOW currents to Alkaline).

NiMH only beat Alkaline on high current draw devices.

Zinc Carbon about 1/5th

EDIT:
BTW the software there is the free Linear Technology "Spice" including schematic capture and loads of modes. Good user group support for it too.

Of course it won't design the circuits, just tell you what they do if the model you draw is realistic.

I felt that the Voltage doubler version should be a bit better than it was...

Click to see full size image

Adding a 4th diode reverse across the switch helps.

I've got the capacitors down in size and peak volts on the IC switch transistor more happily below 50V. Using similar filter and same Switching inductor.

It's 77V steady state, but the end point of Dry Cell B126 original is 54V. The Invicta 29 runs well at 72V on battery though on Mains it's 104V

Click to see full size image

Red is the voltage at feedback point, thin line is voltage at radio and blue is voltage across the LT1073 switch transistor.

I halted simulation as after about 75ms it's stable.

So if 4 x high speed 1A diodes are cheaper than 1 x high speed diode + MOSFET, then it's a feasible solution. The "strange bit" at start up is current limiting activated. (The 68 Ohm R3).

I'll build both and compare at 1mA to 10mA load. Typical Dx96 set should be 5mA to 7mA I think.

I built this using 30 cent LM393 dual comparator.

With a cheap BC547 Bipolar transistor it's OK up to about 50V and will run with 1V supply.

Using a MOSFET needs 5V to 7V supply minimum depending on FET. It will do 900V from 12V with suitable FET and Diodes

Anyway, this one was really running on my workbench a few months ago, but the regulation vs input voltage is poor as it uses a zener as reference. The Zener has to be a lower voltage than supply volts. Below 5.6V Zeners are very soft curve. Hence this uses a 6.2V Zener.

Click to see full size image

Output voltage is set by R1 & R2 where R2 is the Zener voltage.

U1 is a classic Relaxation oscillator. Like an RC with a neon or unijunction transistor. If the output is too high vs the Zener Reference the U2 "shorts" gate or base to 0V. The comparators can be conventional open collector with no pull up.

I never modelled it in Spice (I just built it out of my head at the time). So I have the LT1017 dual comparator simply because it was the 1st suitable one I found on the list of models.

Click to see full size image

Again thin line is filtered volts on Radio, Red is voltage at Feedback point (R1, R2) and Blue is the voltage on FET Drain. The output voltage is stable only 34ms after power on.

68uH is a high current type (1A to 2A rating!) even though average is under 100mA.
100uH is only got 5mA to 10mA so is a cheap miniature resistor type.

I built one with two transistors as multi-vibrator and also a 555 Timer based one. The LM393 dual comparator was best.

But in the end for good low voltage input regulation (1V supply) I thought I'd get some LT1073.

You can run a 12V drive FET off 1.5V battery by making a 12V PSU with no external switch from one LT1073 and use it as PSU just for the 2nd LT1073 drive. The Coil and FET drain then from the 1.5V supply.  But below 50V from 1 to 3 AA or other single cells, a bipolar transistor is simpler.

EDIT
Yahoo LT Spice group http://tech.groups.yahoo.com/group/LTspice/